LINQ - Removing items in a List that contain one item of an Array in any order or position

Question

I am trying to write something similar to the following with LINQ:

var media = from s in db.Media select s;       
string[] criteria = {"zombies", "horror"};

mediaList.RemoveAll(media.Where(s => s.description.Inersect(criteria).Any())); 

//mediaList is a List(T) containing instances of the Media model.

I thought linq where list contains any in list's solution would apply in this case but my compiler complains that "string does not contain a definition for Intersect".

The behaviour I am expecting is for Media items that contain the words zombies or horror but not both in their description to be taken out of the list i.e.

A horror movie.
A movie with a lot of zombies.

But items like the following should stay in the list:

A horror movie with zombies.
The best zombies and the best horror.

The Media class:

public class Media
{
    public int mediaID { get; set; }
    public string type { get; set; }
    public string description { get; set; }
}

The description field contains very long paragraphs. I am afraid the solution is very obvious but for the life of me I cannot work it out.

EDIT: added a better explanation of the behaviour expected.

René Vogt · Accepted Answer · 2019-09-03T16:57:32.577

3

Your confusing some methods here.

List<T>.RemoveAll() takes a Predicate<T> as parameter and removes all elements from the list for which this prediate returns true. So what you want could be somehting like that:

mediaList.RemoveAll(m => criteria.Any(crit => m.description.Contains(crit));

But note that this will also remove "A movie about nonzombies".

UPDATE after your clarification:

mediaList.RemoveAll(m => 
    {
        int count = criteria.Count(crit => m.description.Contains(crit));
        return count > 0 && count < criteria.Length;
    });

This removes all entries that contain at least one word of criteria, but not all of them. (it still does not match "whole words only", though).

edited Sep 03 '19 at 16:57

answered Aug 07 '17 at 13:55

René Vogt

43,056
14
77
99

Apologies but I made some pretty critical mistakes in my question. Particularly on the "behaviour expected" part. – dlp_dev Aug 07 '17 at 14:23
@DanL updated my answer. Please try not to post moving targets. – René Vogt Aug 07 '17 at 14:31
I've been playing around with that code but it always fails at media.Description. The IntelliSense does not show any of the Media properties after I type media and instead shows "string does not contain a definition for description". – dlp_dev Aug 07 '17 at 14:59
@DanL I fixed the spelling (`description` is lower case in your code). But the error you describe is strange. In my second code sample, _if_ `mediaList` really is a `List`, then `media` is of type `Media`, not `string`. Please verify that you copied the code correctly. – René Vogt Aug 07 '17 at 15:04
@DanL Note that `Intersect` as in your code can't work as `description` is a `string`, not an array of "words". – René Vogt Aug 07 '17 at 15:05
made a little mistake while copying although the error is still the same.`mediaList` is of type `List` but seems like the `media.description` is using `media` instead - which was declared on the first line of my code and is a `IQueryable`. – dlp_dev Aug 07 '17 at 15:27
Also note that the local cannot me called `media` as it throws _cannot be declared in this scope because that name is used in an enclosing local scope to define a local or parameter_. (I used it on `var media = from s in db.Media select s; `) Easily fixed by just naming it something else. – dlp_dev Aug 07 '17 at 15:30
@DanL you're right, `media` was already used. I renamed it to `m` in my answer. Did that solve the problem? – René Vogt Aug 07 '17 at 15:32
I'm afraid it still throws a _IQueryable does not contain a definition for 'description'_ – dlp_dev Aug 07 '17 at 15:35
Then you still use the `media` variable name, be sure to replace it inside the `RemoveAll(....)` method call/lambda. – René Vogt Aug 07 '17 at 15:37
You are correct, although when running the code `crit` throws a _NullPointerException_ inside the `.Contains()`. Sorry to keep bugging you. – dlp_dev Aug 07 '17 at 15:49
1

So you probably have rows in your db where `description` is `NULL`. you can fix that for example via `m.description?.Contains(crit) ?? false`... but that's another (basic c#) question. I leave my office now, so this will be my last comment on that. Hope my answer could help so far. – René Vogt Aug 07 '17 at 15:52
Incredibly helpful. Thanks! – dlp_dev Aug 07 '17 at 15:53
@RenéVogt: what's the `&& count < criteria.Length` about? It makes sure you never match 3 out of 2 criteria. The `m => criteria.Any(crit => m.description.Contains(crit)` predicate was sufficient. – tinudu Aug 07 '17 at 17:25
@tinudu no, it makes sure you don't match 2 out of 2 criteria, which was Op's later requirement. – René Vogt Aug 07 '17 at 17:42

score 1 · Answer 2 · answered Aug 07 '17 at 13:56

1

You should use

  var reuslt = mediaList.RemoveAll(media => criteria.Any(c => s.description.Contains(c));

answered Aug 07 '17 at 13:56

Sajeetharan

216,225
63
350
396

score 0 · Answer 3 · answered Aug 07 '17 at 14:12

You can't intersect a string with a string[] array, but you could split the description string into words and then do the intersection:

mediaList.RemoveAll(entry => entry.description.Split(new string[]{" "},
   StringSplitOptions.None).Intersect(criteria).Any());

This avoids the problem of matching words that are containing a substring of the criterion strings.

LINQ - Removing items in a List that contain one item of an Array in any order or position

3 Answers3